Demand response (DR) is the reduction in a customer’s electric load during periods of peak demand or high market prices. The most basic DR programs are structured to maintain system reliability and avoid blackouts and brownouts. In recent years, DR has evolved into a more dynamic resource, providing price mitigation and ancillary services in addition to maintaining reliability. DR is employed to offset or defer the need for generation, transmission and distribution infrastructure.
Policy makers, utilities and system operators have become increasingly interested in DR as a cost-effective, environmentally sensible means to meet system needs. These stakeholders are tasked with designing a successful DR program that meets the specific objectives for each unique system. This paper provides a framework for designing a dispatchable DR program for commercial, institutional and industrial (C&I) end-users.
Designing a successful DR program is a balancing act. Several key parameters must be adjusted to achieve the objectives of the utility or system operator while accommodating the capabilities and constraints of end-use customers. As Figure 1 below suggests, ideal DR program design considers both the needs of the utility or system operator and C&I customers to maximize the potential DR resource.
The following are seven major program design elements to balance when creating a DR program: program compensation, performance measurement, resource response time, resource availability, dispatch triggers, non-performance penalties and program administration. A successful DR program design effectively balances these seven elements to maximize the amount of capacity and number of C&I end-users, all while meeting the needs of utilities and system operators.
DR program participation is influenced by the level and structure of compensation to C&I customers. In general, higher payments increase willingness for more frequent customer demand reductions. Compensation includes capacity payments ($/kW-month), availability/reservations payments ($/ kW-hour) and/or energy payments ($/kWh). Table 1 shows the different types of payments typically included as part of a DR program to compensate the three different types of electricity products — capacity, ancillary services and energy.
For example, in the PJM Interconnection (PJM), DR resources in the Emergency Load Response Program receive monthly capacity payments as well as energy payments during events; PJM’s Synchronized Reserve Market pays users an hourly availability payment if reserved and an energy payment during reductions; PJM’s Economic Load Response Program pays end users a locational marginal price for reductions in the Real-Time or Day-Ahead Energy Markets.
Establishing an accurate measurement and verification methodology for DR performance is critical for the success of DR as a trusted resource. The majority of today’s DR programs use a baseline methodology that relies on recent or historical data to determine what a customer’s load would have been, absent participation in a DR event. The most accurate baseline methodologies incorporate an adjustment factor based on the metered customer’s load on the day of a DR event to capture the effects of weather or other conditions that influence a customer’s demand. There have been a number of academic and consultant studies conducted to determine which baseline methodology provides the most accurate estimates of performance and minimizes systematic bias in either the positive or negative direction (Goldberg & Agnew 2003; Quantum Consulting 2004). To determine the best baseline methodology to use for a DR program, be sure to factor in the types of customers likely to be enrolled in the program and the complexity and timeliness of the baseline calculation. Customers with demand that correlates closely with outside temperatures are more accurately measured using a baseline methodology that explicitly accounts for temperature (Coughlin et al. 2008). Other customers that have relatively stable electric demand may be just as accurately measured using a simple averaging baseline methodology. Ensuring that a customer is able to understand how their performance is being measured and receive feedback (and payment) on performance in a timely manner can contribute to the success of a DR program.
Resource Response Time
Resource response time refers to the amount of time the DR resource has to curtail after being notified of a DR event. Programs can be categorized into either day-ahead or day-of notification. For day-of programs, the response time can vary from several hours to minutes. DR programs designed for participation as ancillary services must be available within 10 minutes or less, depending on the service provided. Many day-of programs provide response times within 30 minutes to 1 hour. In general, there are increased technological requirements for programs with short response times, including very granular metering requirements (e.g., 1-minute) and direct load control capabilities. In addition, the shorter the response time, the fewer the number of customers who will be able to participate in the program. Utility and system operators must balance their interest in having a quick- response resource with their objective of having widespread DR program participation.
Utilities and system operators must decide when the DR resources will be available to respond to dispatch. Resource availability can vary by 1) season; 2) days of the week; 3) hours of the day; and 4) length of the event. DR programs that are designed to achieve peak load reductions for a summer-peaking utility may only be needed for five months of the year. Alternatively, a dual-peaking system may need DR to be available year-round to address all the critical system peak hours. Some DR programs operate only on weekdays and non-holidays; others operate seven days per week. Similarly, the window of availability of a DR program during a given day can be tailored to address the specific needs of a system. For example, DR resources might only be available from noon until 6pm or they might be available 24-hours per day. Finally, the period of time for which a program must operate when called upon can vary from minutes in the case of ancillary services, to hours in the case of reliability and economic DR programs. The selected periods of availability and operation can have significant implications for the overall cost and amount of DR resources available to the program.
DR programs can be dispatched based on a number of different system factors or “triggers.” For example, triggers can include official “grid emergencies,” reserve margins falling below a certain percentage, energy market prices at or above a certain level, voltage reductions, marginal generation heat rates and local distribution emergencies. Some DR programs have limited triggers for which they can be dispatched whereas other DR programs can be dispatched at the full discretion of the utility or system operator. To maximize DR performance, it is helpful if participants understand the dispatch trigger.
Utilities and system operators are increasingly incorporating penalties for non-performance into DR contracts and programs to ensure the reliability of the resource. Alternatively, contracts and programs can have provisions that reduce or eliminate program payments as a result of poor performance. Substantial non-performance penalties can provide strong incentives for reliable DR program performance. However, penalties that are set excessively high can result in poor DR penetration. Third- party DR providers can help shield participating customers from this risk by absorbing penalties.
DR can be administered either by utilities directly or by third-party DR providers. Historically, utilities have administered load management programs directly with their customers. However in recent years, new models emerged, creating new opportunities to integrate DR into the resource mix. In the early 2000s, independent system operators in the eastern United States (ISO New England, the New York Independent System Operator and PJM) constructed a new type of market participant to represent demand response resources in the market place. These entities are now known by various terms including demand response provider, curtailment service provider and responsible interface party. Today, some of the same third-party entities that successfully brought DR to deregulated regions have expanded throughout North America to provide outsourced DR programs to utilities. Utilities are increasingly contracting with third-parties to complement an existing demand side management portfolio with a targeted, dispatchable DR program for C&I or residential consumers. Figure 2 portrays an illustrative load aggregation model employed by third-party DR providers, where the DR provider manages a portfolio of customers to provide the utility or grid operator with a specified level of DR capacity during DR events. In this model, the DR provider absorbs the program risk and shields the customer from potential non-performance penalties, while providing the grid operator or utility with the expected DR resource.
This section provides a framework on balancing the key DR program design elements described above.
Step 1: Define Program goals/Objectives
Given multiple design decisions, the first step in the development of any DR program is to clearly define the goals of the program and the measures of success. For example, one utility may wish to develop a DR program exclusively for reliability purposes during peak summer months; a successful program would meet reserve obligations with DR while avoiding blackouts/brownouts. Alternatively, a second utility may develop a DR program to meet resource needs where DR is more economic rather than building or buying alternative supply-side resources; this program is successful if it yields a more economically-efficient system through DR dispatch.
Step 2: Balance Key Program Design Elements to Best Achieve goals/Objectives
Figure 3 shows the range of choices for DR program design elements and how the selection made can help or hinder the resource potential. The obvious example is program compensation with higher incentive payments to customers, this DR program attracts more customers and better performance. However, higher program compensation leads to higher program costs. Therefore, the DR program designer must adjust the other design parameters to create a high-value and cost-effective resource. Decisions surrounding many design elements should depend largely on the goals of the program. For example, if the goal of the program is to supply ancillary services, a one to ten minute response time might be needed to achieve the goal; on the other hand, if the goal of the DR program is to improve the system reliability, then system operators might be able to effectively manage a resource with a two hour response time.
Step 3: Consider Unintended Consequences
When designing a new DR program, it is important to consider how it will interact with existing DR programs. For example, if many of a utility’s larger customers are on an existing interruptible tariff, then introducing a DR program that appeals to those same customers may simply result in a shift of DR resources rather than an expansion of the total pool of available DR. In essence, designers should avoid cannibalizing existing (and successful) programs and instead seek out non-participating customers through complimentary DR programs. If a new DR program has too many points of overlap for existing and successful programs, consider revising some of the key program elements to make the new program unique and targeted. Alternatively, the existing program could be updated with changes to program elements to expand its size or improve upon its proven success.
Step 4: Finalize Program
Once all the key program design elements are determined, the remaining operational aspects of the program can be finalized. In Table 2, several DR program goals are listed across the top and the elements of design listed along the left-hand side. The ranges for the design elements suggest that even with the goal of the DR program identified, there is further room for refinement in design. It is important for the utility or grid operator to avoid “over-designing” a single DR program by trying to capture all possible scenarios for which the DR resource may or may not be needed. This actually has the effect of limiting the program’s ability to meet any of the intended scenarios. A better approach is to design multiple DR programs, each with their own goals and success factors, which can be used as part of a targeted resource strategy. Additional program finalization steps might include establishing expectations around the frequency of DR events; determining metering requirements, communications protocols and requirements; and developing marketing and outreach strategies. Most programs also require regulatory or board approval of the new program.
Step 5: Keep an Eye Out for Future Enhancement Opportunities
Just as previous generations of load management programs have evolved over the last several years, it is clear that today’s programs will benefit from future technology innovations and changes in the electric market structure. For example, the effect of improved metering infrastructure undoubtedly has an impact on the next generation of DR. Consideration of the positive environmental impact DR can have in future climate change mitigation efforts will need to be evaluated as part of DR program design.
One of the more immediate considerations for DR program designs is the potential for integration of DR and other demand-side resources, such as energy efficiency and distributed generation. In both California and New York, policy-makers are evaluating the combined effect of DR and energy efficiency resources. One evaluation of the environmental benefits of DR and energy efficiency asserts that “…consumers rarely are interested in the distinctions among demand-side measures... but rather in bottom-line results — lower power bills, rebates on new equipment, lessened risk and better environmental performance. Energy efficiency and DR advocates may well find that working together to promote overall demand- side management may yield political results that could not be achieved by either side alone.” (Nemtzow, Delurey & King, 2007)
DR has evolved significantly in the past several years as a result of improved enabling, metering and communication technologies, utilities’ increased need to manage peak demand amidst rising infrastructure and resource costs, and the emergence of competitive demand response suppliers. For the first time, DR is effectively and efficiently incorporated into electric system resource planning or grid management strategies as an alternative to new peaking capacity or transmission and distribution infrastructure. However, in order for DR resources to achieve these objectives, the programs need to continue to improve over time. Designing a successful DR program is key to enabling the full value of the DR resource now and in the future.
Coughlin, Katie, Mary Ann Piette, Charles Goldman, and Sila Kiliccote, 2008. Estimating Demand Response Load Impacts: Evaluation of Baseline Load Models for Non- Residential Buildings in California, Demand Response Research Center, Lawrence Berkeley National Laboratory.
Goldberg, Miriam L. and G. Kennedy Agnew 2003. Protocol Development for Demand-Response calculations: Findings and Recommendations. Prepared for the California Energy Commission by KEMA Xenergy. CEC 400-02-017F
Nemtzow, D., D. Delurey, and C. King, The Green Effect: How demand response programs contribute to energy efficiency and environmental quality, Public Utilities Fortnightly, March 2007: 45.
Quantum Consulting Inc. and Summit Blue Consulting, LLC, 2004. Working Group 2 Demand Response Program Evaluation — Program Year 2004 Final Report, Prepared for the California Working Group 2 Measurement and Evaluation Committee.